ABSTRACT

Structural health monitoring has been developed rapidly for jacket platform during the latest decades. However, the changes of the structural dynamic responses do not depend only on the damage but also on ocean environmental conditions. The PCA-based method has been proved effective in discovering the hidden information from grossed data. Additionally, the cross-correlation function has been proved effective in building structural characteristics by proceeding it between the dynamic responses from different structural positions. Thus, this paper presents an approach in developing a PCA-based method for clarifying the structural health state from the structural characteristics built by cross-correlation function.

INTRODUCTION

The offshore oil industry grows rapidly during recent decades and plays an important role in the global oil industry. Steel jacket platforms are the most common used structures in exploring oil or gas resources from ocean (Wisch, 1998). Presently, the design, construction and installation technologies can provide sufficient safety guarantee for the operation of jacket platforms. However, unexpected structural damage may occur during their service lives and probably lead the structure into a dangerous situation. Therefore, effective, reliable, and real-time structural health monitoring (SHM) is important and necessary to jacket platforms (Nichols, 2003; Roitman, Gadea and Magluta, 2004).

The aim of the SHM is to detect, locate and assess the extent of damage so that the structure's service lives can be possibly extended. The general method for SHM is to extract meaningful features from the measured data. These features are monitored in order to detect changes due to damage. At present, there have been large amounts of literatures published on extraction of features. For example, the natural frequency, mode shape, mode shape curvature and modal flexibility are commonly used as the damage-sensitive features (Carden and Fanning, 2004). For offshore platform, the above modal-based parameters are difficult to extract precisely due to the excitations influenced by wave loads. To avoid the limitation of the modal parameters, some researchers detect the damage by using the auto/cross-correlation functions, coherence functions, and the analysis on the vibration signal with time series analysis.

This content is only available via PDF.
You can access this article if you purchase or spend a download.